Scolaris Content Display Scolaris Content Display

Laparoscopic versus open total mesorectal excision for rectal cancer

Collapse all Expand all

Background

Colorectal cancer including rectal cancer is the third most common cause of cancer deaths in the western world. For colon carcinoma, laparoscopic surgery is proven to result in faster postoperative recovery, fewer complications and better cosmetic results with equal oncologic results. These short‐term benefits are expected to be similar for laparoscopic rectal cancer surgery. However, the oncological safety of laparoscopic surgery for rectal cancer remained controversial due to the lack of definitive long‐term results. Thus, the expected short‐term benefits can only be of interest when oncological results are at least equal.

Objectives

To evaluate the differences in short‐ and long‐term results after elective laparoscopic total mesorectal excision (LTME) for the resection of rectal cancer compared with open total mesorectal excision (OTME).

Search methods

We searched the Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library 2013, Issue 2), MEDLINE (January 1990 to February 2013), EMBASE (January 1990 to February 2013), ClinicalTrials.gov (February 2013) and Current Controlled Trials (February 2013). We handsearched the reference lists of the included articles for missed studies.

Selection criteria

Only randomised controlled trials (RCTs) comparing LTME and OTME, reporting at least one of our outcome measures, was considered for inclusion.

Data collection and analysis

Two authors independently assessed study quality according to the CONSORT statement, and resolved disagreements by discussion. We rated the quality of the evidence using GRADE methods.

Main results

We identified 45 references out of 953 search results, of which 14 studies met the inclusion criteria involving 3528 rectal cancer patients. We did not consider the risk of bias of the included studies to have impacted on the quality of the evidence. Data were analysed according to an intention‐to‐treat principle with a mean conversion rate of 14.5% (range 0% to 35%) in the laparoscopic group.
There was moderate quality evidence that laparoscopic and open TME had similar effects on five‐year disease‐free survival (OR 1.02; 95% CI 0.76 to1.38, 4 studies, N = 943). The estimated effects of laparoscopic and open TME on local recurrence and overall survival were similar, although confidence intervals were wide, both with moderate quality evidence (local recurrence: OR 0.89; 95% CI 0.57 to1.39 and overall survival rate: OR 1.15; 95% CI 0.87 to1.52). There was moderate to high quality evidence that the number of resected lymph nodes and surgical margins were similar between the two groups.
For the short‐term results, length of hospital stay was reduced by two days (95% CI ‐3.22 to ‐1.10), moderate quality evidence), and the time to first defecation was shorter in the LTME group (‐0.86 days; 95% CI ‐1.17 to ‐0.54). There was moderate quality evidence that 30 days morbidity were similar in both groups (OR 0.94; 95% CI 0.8 to 1.1). There were fewer wound infections (OR 0.68; 95% CI 0.50 to 0.93) and fewer bleeding complications (OR 0.30; 95% CI 0.10 to 0.93) in the LTME group.
There was no clear evidence of any differences in quality of life after LTME or OTME regarding functional recovery, bladder and sexual function. The costs were higher for LTME with differences up to GBP 2000 for direct costs only.

Authors' conclusions

We have found moderate quality evidence that laparoscopic total mesorectal excision (TME) has similar effects to open TME on long term survival outcomes for the treatment of rectal cancer. The quality of the evidence was downgraded due to imprecision and further research could impact on our confidence in this result. There is moderate quality evidence that it leads to better short‐term post‐surgical outcomes in terms of recovery for non‐locally advanced rectal cancer. Currently results are consistent in showing a similar disease‐free survival and overall survival, and for recurrences after at least three years and up to 10 years, although due to imprecision we cannot rule out superiority of either approach. We await long‐term data from a number of ongoing and recently completed studies to contribute to a more robust analysis of long‐term disease free, overall survival and local recurrence.

PICOs

Population
Intervention
Comparison
Outcome

The PICO model is widely used and taught in evidence-based health care as a strategy for formulating questions and search strategies and for characterizing clinical studies or meta-analyses. PICO stands for four different potential components of a clinical question: Patient, Population or Problem; Intervention; Comparison; Outcome.

See more on using PICO in the Cochrane Handbook.

Plain language summary

Keyhole laparoscopic or open surgery for rectal cancer

Colorectal (large bowel) cancer including rectal cancer is the third most common cause of cancer deaths in the western world. The risk of developing rectal cancer increases with age and is most common in people around 70 years of age. The treatment consists of complete surgical resection of the tumour and surrounding tissue by a technique called total mesorectal excision (TME), sometimes combined with chemotherapy and radiotherapy. This surgery can be performed by either normal open abdominal surgery with a large incision or by keyhole laparoscopic surgery with several small incisions for the instruments and camera. For colon cancer, laparoscopic surgery is proven to result in faster postoperative recovery, fewer complications and better cosmetic results. These results are expected to be equal for rectal surgery. However, surgery for rectal cancer is technically more difficult than for colon cancer due to the location deeper in the pelvis and close to important nerves. Therefore a complete and safe resection of the tumour should be guaranteed, this is important to reduce the risk of recurrence of the tumour and could be tested by assessing recurrence rates and patient survival in the long term.

In this updated review, we have assessed all randomised studies of laparoscopic and open TME for rectal cancer, to compare and combine their results. We included 14 trials reporting on a total of 3528 patients undergoing rectal cancer surgery. In 14.5% of those having laparoscopic surgery needed conversion to open surgery by a large incision in the abdomen due to difficulties or problems during the procedure.

There is currently moderate quality evidence that laparoscopic total mesorectal excision (LTME) has similar effects to open TME (OTME) on long term survival outcomes for the treatment of rectal cancer. The estimated effect was imprecise and further research could impact on our confidence int this result. There is moderate quality evidence that it leads to better short‐term post‐surgical outcomes in terms of length of hospital stay. We found that pain was lower in the LTME group and that resumption of diet was better. We did not find clear evidence of a difference in quality of life between the two groups, but costs were higher for LTME. We await long‐term data from a number of ongoing and recently completed studies to contribute to our understanding of the effects of these surgical approaches on long‐term disease free, overall survival and local recurrence.

Authors' conclusions

Implications for practice

There is currently moderate quality evidence that laparoscopic total mesorectal excision (TME) has similar effects to open TME on long term survival outcomes for the treatment of rectal cancer. The quality of the evidence was downgraded due to imprecision and we cannot rule out either approach being superior. There is moderate quality evidence that it leads to better short‐term post‐surgical outcomes in terms of recovery for non‐locally advanced rectal cancer and shorter hospital stay. Currently results are consistent in showing a similar disease‐free survival and overall survival, and for recurrences after at least three years and up to 10 years although due to imprecision we cannot rule out superiority of either approach. We await long‐term data from a number of ongoing and recently completed studies to contribute to a more robust analysis of long‐term disease free, overall survival and local recurrence.

Implications for research

The evidence presented in this systematic review is sufficient to establish the overall and long‐term oncological safety of laparoscopic TME. However, at this moment the available data are still insufficient to confirm these results for subgroups such as abdominoperineal resections, following neoadjuvant therapy, in locally advanced disease and in combination with fast track recovery protocols.

Summary of findings

Open in table viewer
Summary of findings for the main comparison. Laparoscopic versus open total mesorectal excision for rectal cancer

Laparoscopic versus open total mesorectal excision (TME) for rectal cancer

Patient or population: people with Rectal Cancer
Settings:
Intervention: Laparoscopic TME
Comparison: Open TME

Outcomes

Illustrative comparative risks* (95% CI)

Relative effect
(95% CI)

No of Participants
(studies)

Quality of the evidence
(GRADE)

Assumed risk

Corresponding risk

Open TME

Laparoscopic TME

Disease‐free survival at 5 years

718 per 1000

722 per 1000
(659 to 778)

OR 1.02
(0.76 to 1.38)

943
(4 studies)

⊕⊕⊕⊝
moderate1

Overall survival at 5 years

679 per 1000

709 per 1000
(648 to 763)

OR 1.15
(0.87 to 1.52)

987
(4 studies)

⊕⊕⊕⊝
moderate2

Local recurrences

54 per 1000

48 per 1000
(31 to 73)

OR 0.89
(0.57 to 1.39)

1538
(8 studies)

⊕⊕⊕⊝
moderate3

Lymph nodes retrieved

The mean number of lymph nodes retrieved in the intervention groups was
0.43 lower
(1.13 lower to 0.26 higher)

3682
(11 studies)

⊕⊕⊕⊕
high

CRM positivity

61 per 1000

60 per 1000
(44 to 83)

OR 0.99
(0.71 to 1.4)

2313
(8 studies)

⊕⊕⊕⊝
moderate4

30‐day morbidity (total)

275 per 1000

263 per 1000
(233 to 295)

OR 0.94
(0.8 to 1.1)

3397
(11 studies)

⊕⊕⊕⊝
moderate5

Hospital stay (days)

The mean length of hospital stay in the intervention groups was
2.16 days shorter
(3.22 to 1.1 days shorter)

3084
(11 studies)

⊕⊕⊕⊝
moderate6

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; OR: Odds ratio;

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

1Statistical inaccuracy with wide confidence interval at both sides

2Statistical inaccuracy with wide confidence interval at both sides, but a tendency for a higher overall survival for LTME

3Statistical inaccuracy with wide confidence interval at both sides, but a tendency for a lower recurrence rate for LTME

4Only 8 studies provided data on CRM positivity
5Definition of overall morbidity varied or was unclear
6Length of hospital stay depends on postoperative protocols and implementation of enhanced recovery programmes

Background

Description of the condition

The incidence of rectal cancer in the western world is 28% to 35% of the total colorectal cancer incidence, with 15 to 25/100,000 new patients per year for both men and women. The risk increases with age, with a median age of 70 years at the time of diagnosis; the associated mortality is between 4 and 10 per 100,000 per year (Glimelius 2013). Symptoms suggesting rectal cancer typically include changes in bowel habits, the feeling of incomplete emptying, rectal bleeding, anaemia or weight loss. The diagnosis can be made by tumour biopsy during colonoscopy or sigmoidoscopy. If rectal cancer is confirmed, the extent of the disease is examined by imaging of the chest and liver for signs of metastases, and magnetic resonance imaging (MRI) of the pelvis and/or endorectal ultrasound (ERUS) are done to determine the degree of rectal wall and mesorectal fascia invasion. The majority of rectal carcinomas are adenocarcinoma (95% to 98%), usually arising from an adenoma (Glimelius 2013; Monson 2013).

Description of the intervention

Complete resection of rectal cancer can be achieved by a sphincter‐preserving anterior resection (AR, rectosigmoid resection) or an abdominoperineal resection (APR). Both had high local recurrence rates until the introduction of the total mesorectal excision (TME) (Heald 1986). Total mesorectal excision achieves a complete removal of the rectum together with its draining lymphatics, and results in low rates of recurrence. Despite the successful introduction of laparoscopic and laparoscopic‐assisted procedures for the resection of colonic cancer, surgeons have been more reluctant to introduce laparoscopic TME due to the technically demanding resection plane.

How the intervention might work

Laparoscopic and laparoscopic‐assisted TME offers several theoretical advantages compared to open resection. It may be associated with less blood loss, faster recovery, early feeding and a lower morbidity rate, as shown in laparoscopic colonic surgery (Braga 2002; Pikarsky 2002). The magnified view of the pelvis afforded by the laparoscope may facilitate identification of the autonomic nerves and thus prevent unintentional injury of these nerves. However, these advantages of LTME are only beneficial to people with rectal cancer when local recurrence and disease‐free survival rates are at least similar to those achieved with OTME.

Why it is important to do this review

The introduction of laparoscopy 20 years ago has caused major changes in colorectal surgery. For benign disease, such as diverticulitis and inflammatory bowel disease, laparoscopy has become the surgical technique of choice for its benefits in recovery, complication rate and cosmetic results. Only recently, sufficient evidence has become available showing laparoscopic surgery is safe for the treatment of colonic cancer. Four large randomised trials (472 to 1076 participants) could not show any differences in quality of resection and long‐term recurrence and survival rates between laparoscopic and open surgery for colon cancer (MRC CLASICC a 2005; COST 2007; COLOR 2009; LAPKON II 2009). Although COLOR 2009 was not able to rule out any difference with their non‐inferiority design, the meta‐analysis by Kuhry 2008 did not show any differences.

Despite the larger number of randomised trials on laparoscopic surgery for colon cancer, there is still limited evidence for long‐ and short‐term outcomes after LTME due to the lack of high quality randomised controlled trials with sufficient follow up. Now the results of more well designed large randomised controlled trials (RCTs) become available, there is a need for a updated systematic review of these results.

Objectives

To evaluate the differences in short‐ and long‐term results after elective laparoscopic total mesorectal excision (LTME) for the resection of rectal cancer compared with open total mesorectal excision (OTME).

Methods

Criteria for considering studies for this review

Types of studies

In contrast to the published protocol and previous version of this review, for this update we have only considered RCTs comparing LTME to OTME, since sufficient RCTs have become available since the publication of the original review. We did not apply any language restrictions.

Types of participants

People with rectal cancer undergoing total mesorectal excision were considered for inclusion. Studies including participants with colorectal cancer are only eligible if the results for those with rectal carcinoma were presented separately.

Types of interventions

These include laparoscopic, laparoscopic‐assisted or open total mesorectal excision as (low) anterior resection or abdominoperineal resection. When a primary anastomosis was constructed, it could either be performed intraperitoneally ('double‐stapled' colorectal anastomosis) or extraperitoneally (hand‐sewn or stapled colorectal anastomosis).

Types of outcome measures

We sought the following outcomes in all included studies:

Primary outcomes

‐ Disease‐free and overall survival

Secondary outcomes

‐ Recurrences (local, wound/port site and distant)
‐ Quality of resection (circumferential margin (CRM) positivity and number of lymph nodes)
‐ Surgical data (surgical time, incision length, conversion rate)
‐ Intraoperative complications, blood loss and transfusion requirement
‐ Postoperative morbidity and mortality (overall morbidity, need for reoperation, anastomotic leakage, wound infection, urinary complications, bleeding, chest infection)
‐ Postoperative pain (use of medication and visual analogue scale (VAS) score)
‐ Gastrointestinal recovery and hospital stay (time to first bowel movement, time to normal diet, length of hospital stay)
‐ Long‐term morbidity (incisional herniae and bowel obstruction)
‐ Quality of life (functional recovery, bladder and sexual function)
‐ Immunologic response
‐ Costs

Search methods for identification of studies

Electronic searches

We followed the recommendations of the Cochrane Colorectal Cancer Group and searched the following bibliographic databases with no language restrictions in order to identify relevant primary studies:

Cochrane Central Register of Controlled Trials (CENTRAL) (January 1990 to February 2013);
MEDLINE (January 1990 to February 2013);
EMBASE (January 1990 to February 2013).

We conducted searches using medical subject headings (MeSH) and free‐text words. The search has been adapted for each database search and is shown in Appendix 1 (CENTRAL), Appendix 2 (MEDLINE) and Appendix 3 (EMBASE).

Searching other resources

We handsearched the reference lists of all selected articles for further relevant studies. There was no language restriction. In addition, we searched for ongoing trials in the ClinicalTrials.gov and the Current Controlled Trials databases.

Data collection and analysis

Selection of studies

Two authors (SV and LP) independently reviewed all abstracts. We retrieved full‐text copies of all studies that potentially met the inclusion criteria based on abstract review. If both authors agreed that a study did not meet the eligibility criteria, we excluded it. If we disagreed, we resolved conflicts by discussion and consensus or by consulting a third member of the review team.

Data extraction and management

We collected data according to the outcomes mentioned above. Each author extracted the data independently from each study and compared them, resolving disagreement by discussion. We used Review Manager 5 (RevMan 5.2) software for statistical analysis, provided by the Cochrane Collaboration. Data not suitable for meta‐analysis is discussed in the results section.

Assessment of risk of bias in included studies

Two authors (SV and LP) have assessed all the selected studies for methodological quality according to the CONSORT Statement 2010, and have summarised the information in the 'Risk of bias' figure (Figure 2). In addition, we have used the Cochrane 'Risk of bias' tool as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Measures of treatment effect

We measured the treatment effect using the mean difference (MD) or standardised mean difference (SMD) for continuous data and the odds ratio (OR) for dichotomous data. Standardised mean differences were only used when the reported units or drugs varied between the studies, for instance the number of doses of analgesia. All outcomes included 95% confidence intervals (CI).

Unit of analysis issues

For randomised controlled trials for this surgical intervention, we would expect only a simple parallel design. The only possible cross‐over would be the conversion of laparoscopy to open surgery. Because worse outcomes for this group can be expected and should be evaluated as a complication of laparoscopic‐intended surgery, all analyses should be performed on the intention‐to‐treat principle.

Dealing with missing data

To avoid missing unpublished studies, we searched clinical trial databases as stated above. We compared reported outcomes to published protocols or to the Methods section of each article. If we found inconsistencies, this is reported in the 'Selective reporting' section of the 'Risk of bias' table. As missing postoperative and follow‐up data are common in surgical studies, we assume a random pattern of missingness.

Assessment of heterogeneity

We used the Cochrane Chi² test (Q‐test) to assess heterogeneity and the I² statistic to estimate the degree of heterogeneity (Higgins 2003). We considered an I² of between 0% and 40% as probably not important, between 30% and 60% as representing moderate heterogeneity, between 50% and 90% as substantial heterogeneity, and between 75% and 100% as considerable heterogeneity (Higgins 2011). We used a fixed‐effect analysis for outcomes with low heterogeneity.

Assessment of reporting biases

We present an overview of all outcomes per study in the table Selective reporting (reporting bias).

Data synthesis

We analysed continuous variables using mean differences with 95% confidence intervals. For dichotomous variables we used odds ratios with 95% confidence intervals. We constructed forest plots, using the Mantel‐Haenszel method (fixed‐ or random‐effects) to combine the outcomes. In case of continuous data presented as median and range, we estimated the mean and standard deviation according to the methods described by Hozo 2005. We generated funnel plots to screen for publication bias. In case of inclusion of an original RCT and the additional publication of a subgroup of participants, we included only the most appropriate subgroup data in the meta‐analyses, to avoid duplication of data.

Subgroup analysis and investigation of heterogeneity

We had planned subgroup analyses for abdominoperineal resection (APR) and anterior resection (AR), and for studies allowing and excluding neoadjuvant therapy. These analyses were not performed because too few studies presented separate data for these groups. However, we plan to explore these subgroups if possible in future updates.

Summary of Findings table

We applied methods developed by the GRADE working group to rate the quality of evidence from RCTs, starting at high quality and downgrading for risk of bias, imprecision, inconsistency, indirectness and publication bias.

We rated the quality of the evidence for the following main outcomes:

  1. Disease‐free survival at 5 years

  2. Overall survival at 5 years

  3. Local recurrences

  4. Lymph nodes retrieved

  5. CRM positivity

  6. 30‐day morbidity

  7. Hospital stay (days)

Results

Description of studies

See: Characteristics of included studies, Characteristics of excluded studies and Characteristics of ongoing studies.

Results of the search

Our searches of the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE and EMBASE identified 90, 253 and 852 results respectively. In addition we handsearched MEDLINE for any missed publications for the included trials. After exclusion of duplicates, we screened 953 references and identified 45 eligible references, as shown in the flow chart (Figure 1).


Study selection flow diagram.

Study selection flow diagram.

Included studies

From the 29 references, we identified two ongoing trials and 14 published clinical trials. Most larger trials have published their results at several stages in different papers. Both ongoing trials (ACTRN12609000663257 and NCT00726622) are still recruiting participants and no results have been published yet. Of the 14 published trials, two (Kang 2010; COLOR 2 a 2013) completed participant recruitment but have not yet published long‐term data on survival. The COLOR 2 trial has a second published paper on a local subgroup, referred to as COLOR 2 b 2011 in this review. The Hong Kong trials are divided into the low rectal cancer group in Ng 2008 and the rectosigmoid group in Hong Kong a 2004, with the last subdivided in four papers because results are published for different subgroups as shown in the Characteristics of included studies. Hong Kong a 2004 is the biggest group, presenting upper rectum and sigmoid data, Hong Kong b 2009 is the upper rectal subgroup and reports 10‐year follow‐up data. Hong Kong c 2000 and Hong Kong d 2003 are both small subgroups of Hong Kong a 2004 and present only short‐term data on immunological response.

The UK MRC CLASICC study is presented across nine papers, with six grouped as MRC CLASICC a 2005, giving respectively the short‐term, three‐year, five‐year and 10‐year results, the costs and an analysis of long term complications of the same participant group. MRC CLASICC c 2001 (one) and MRC CLASICC b 2005 (two) include papers for a selected or local participant subgroup and are therefore reported separately. King 2006 also consists of two papers, with the second paper focusing on functional recovery in the same participant group. The remaining eight clinical trials have one reference each, with the note that Zhou 2004 and Zhou 2007 are different trials and different authors despite the coincidence of names.

All 14 clinical trials were published as full papers and involved a total of 3528 rectal cancer patients (range 19 to 1044). The characteristics of these trials are described in 20 separate data sets and thus tables to allow for sufficient details on six additional subgroup papers.

All studies had quite similar exclusion criteria. The most common were: T4 rectal cancer, rectal cancer recurrence, people with synchronous or metachronous colorectal cancer, metastatic disease, emergency surgery, intestinal obstruction or perforation, contraindication for laparoscopy and no informed consent. The majority of the studies described the technique for laparoscopic total mesorectal excision (TME). Perioperative treatment of participants was not described in most of the trials. Six studies had a standardised protocol (Hong Kong a 2004; Braga 2007; Ng 2008; Kang 2010; COLOR 2 b 2011; Liang 2011) and only two had an enhanced recovery protocol (King 2006; Lujan 2009). Data on the type of anaesthesia and analgesia were not given in most studies.

Most studies reported on a range of different outcomes. The most commonly assessed were overall and disease‐free survival rates, local recurrence rates, adequacy of oncological resection (margins and number of lymph nodes removed), duration of surgery, conversion rate, mortality, morbidity, anastomotic leakage, postoperative pain, gastrointestinal recovery rate and hospital stay. Most studies lacked a definition of conversion. The most common causes for conversion to open surgery were tumour invasion of adjacent structures or bulky tumours, dense adhesions and technical failure. Few studies evaluated quality of life (MRC CLASICC a 2005; MRC CLASICC b 2005; King 2006; Braga 2007; Kang 2010), immune response (Hong Kong c 2000; MRC CLASICC c 2001; Hong Kong d 2003; Zhou 2007; COLOR 2 b 2011) or costs (Hong Kong a 2004; MRC CLASICC a 2005; King 2006; Braga 2007; Ng 2008).

Excluded studies

Sixteen papers were excluded for the following reasons: two studies were not completely randomised (Leung 1999; Mirza 2008) and two other studies presented the same data as another included study (Braga 2002; Braga 2005). Two studies included participants with benign disease (Milsom 1998; Polle 2007) and two other studies excluded people treated with TME for low rectal cancer (Schwenk 1998; Liu 2009). Four studies included colorectal cancer patients, but did not report the number of rectal cancer patients or any separate outcomes for rectal cancer (Kim 1998; JCOG 0404 2005; LAPKON II 2009; LaFa 2011). One study presented low‐quality data from a period with a steep learning curve (Pan 2007). Three more studies turned out not to be prospective RCTs, but were a comparison with the national registry (Morris 2011), an economic comparison between UK and USA trials (Stead 2000), and a single‐arm phase II trial (Yamamoto 2008).

Risk of bias in included studies

The risk of bias is described in the Characteristics of included studies section, and a summary is shown in Figure 2. Of the included trials, only one had a low risk of bias on all items and three scored low on six out of seven domains. Five were of unclear or low quality, with a high or unclear risk in at least five out of seven domains.


Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

The method of randomisation was unclear in four trials, and allocation concealment was not described in seven trials. Only seven trials presented an adequate inclusion and randomisation flow diagram, including a description of the loss to follow‐up.

Blinding

Because of the nature of the interventions, blinding is not an option in these trials. Instead of blinding, we assessed whether operative technique and postoperative care were standardised, and how outcome data and pathological data were registered. We assessed standardisation in three studies as inadequate. Outcome registration was adequate in eight studies and unknown in all other studies.

Incomplete outcome data

We did not detect any attrition bias. Not all studies reported on loss to follow‐up. Questionnaire response rates were reported for both groups when eligible.

Selective reporting

Comparing the described protocols and methods to the reported results of the different studies, we did not find evidence of any selective reporting, although some studies did not report exact data on non‐significant results mentioned in the text sections. An overview of studies and outcomes in Table 1 shows that most studies report on the same outcomes. Five papers reported only one outcome in a subgroup analysis of another trial and one study reported only one outcome for the included participant group.

Open in table viewer
Table 1. Reported outcomes

Study ID

n

Long‐term survival

30‐day mortality

30‐day morbidity

Long‐term morbidity

Lymphnodes

Gastrointestinal recovery

Pain

Bleeding

Length of hospital stay

Immune response

Quality of life

Cost

Araujo 2003

28

+

+

+

+

Braga 2007

168

5y/3y

+

+

+

+

+

+

+

+

+

COLOR 2 a 2013

1044

+

+

+

+

+

+

+

COLOR 2 b 2011

40

+

+

+

+

+

+

Hong Kong a 2004

403

5y

+

+

+

+

+

+

+

+

Hong Kong b 2009

153

10y

+

+

Hong Kong c 2000

34

+

Hong Kong d 2003

40

+

Kang 2010

340

+

+

+

+

+

+

+

+

King 2006

19

+

+

+

+

+

+

Liang 2011

343

3y

+

+

+

+

+

Liu 2010

186

+

+

+

+

+

Lujan 2009

204

5y

+

+

+

+

+

+

MRC CLASICC a 2005

381

10y/5y/3y

+

+

+

+

+

+

MRC CLASICC b 2005

148

+

MRC CLASICC c 2001

236

+

Ng 2008

99

5y

+

+

+

+

+

+

+

+

Pechlivanides 2007

73

+

Zhou 2004

171

+

+

+

+

+

Zhou 2007

71

+

Other potential sources of bias

An important source of bias is the experience of the surgeons conducting LTME, because of the known steep learning curve. (Schlachta 2001; Tekkis 2005) When only one experienced surgical team is involved, this bias can be limited, but it can be extensive in large multi centre trials or less experienced teams. Only the MRC CLASICC a 2005 and Kang 2010 defined the experience of their surgeons as based on at least 20 procedures. Four other studies only stated that their surgeons were "well experienced" or "the most experienced" (Hong Kong a 2004; Braga 2007; Pechlivanides 2007; Ng 2008;). The remaining eight did not describe surgeons' experience at all or stated only a single surgeon or team performed the procedures.

Effects of interventions

See: Summary of findings for the main comparison Laparoscopic versus open total mesorectal excision for rectal cancer

Disease‐free and overall survival
The disease‐free and overall survival rates have been reported in only six studies including 1494 participants, because of lack of follow‐up in the other eight studies. Two of these are trials that will report on these results in the near future (Kang 2010; COLOR 2 b 2011), while the other six did not mention any long‐term outcomes in their Methods or protocol.

The combined data for these studies do not show statistical significant differences in disease‐free survival at three ((OR 1.08; 95% CI 0.67 to 1.74), five (OR 1.02; 95% CI 0.76 to 1.38) and 10 years ((OR 1.25; 95% CI 0.51 to 3.06) Analysis 1.1) for LTME and OTME. Regarding overall survival at three (OR 1.00; 95% CI 0.70 to 1.42), five (OR 1.15; 95% CI 0.87 to 1.52) or 10 years ((OR 1.15; 95% CI 0.80 to 1.65); Analysis 1.2), again no differences could be found between the groups. Braga 2007 could not be included in the meta‐analysis because data were only shown in a Kaplan‐Meier curve, but did not show any differences between LTME and OTME groups. See Figure 3.


Forest plot of comparison: 2 Survival and recurrences, outcome: 2.1 Disease free survival.

Forest plot of comparison: 2 Survival and recurrences, outcome: 2.1 Disease free survival.

Ng 2012 has reported the combined 10‐year follow‐up of Hong Kong a 2004 and Ng 2008 (n = 278) in a conference abstract, and reported no statistically significant differences in survival and recurrences (disease‐free survival 82.5% versus 77.6%, P = 0.443, overall survival 63.0% versus 61.1%, P = 0.505 and locoregional recurrences 5.5% versus 9.3%, P = 0.296).

Recurrences
There are no statistical significant differences seen in recurrence rates between LTME and OTME (local OR 0.89; 95% CI 0.57 to 1.39; Analysis 1.3, and distant OR 0.96; 95% CI 0.70 to 1.32; Analysis 1.4). As for port site metastases, only 11 participants (0.9%) in the LTME group developed a port site metastasis (Analysis 1.5). Eight were an extraction site recurrence, leaving only one true port site metastasis. The other two recurrences were not specified. Only three studies described the use of a wound protector in LTME, and reported a similar total of two (0.8%) port site recurrences (Hong Kong a 2004; Zhou 2004; Kang 2010).

Quality of resection: CRM positivity and number of lymph nodes retrieved
One of the most important variables for measuring the quality of the oncological resection and predicting recurrence and survival are circumferential margin involvement and number of lymph nodes retrieved. Eleven RCTs describe the number of retrieved lymph nodes, with no difference between both groups (MD ‐0.43; 95% CI ‐1.13 to 0.26; Analysis 2.1). Eight studies reported on circumferential margin positivity, with no difference between LTME and OTME (OR 0.99; 95% CI 0.71 to 1.40; Analysis 2.2).

Duration of surgery, incision length and conversion rate
The duration of surgery was longer for LTME in 11 out of 12 included trials for this analysis, with a difference of 37 minutes (MD 37.48; 95% CI 27.80 to 47.15; Analysis 2.3). Two other studies did not report on surgical time. Four studies reporting on incision length all found a shorter incision length for LTME with a mean difference of 12 centimetres (MD ‐12.83; 95% CI ‐14.87 to ‐10.80; Analysis 2.4).

All studies except two (Zhou 2004; Zhou 2007) describe the conversion rate for the laparoscopic group (Analysis 2.5). The mean conversion rate was 14.5% (0% ‐ 34%), and as described in MRC CLASICC a 2005, is highly dependent on the location of the tumour and the experience of the surgeon. For most studies, the surgeons' experience was not clearly stated and therefore cannot be compared from these results.

Intraoperative morbidity, blood loss and transfusion requirement
Ten studies described less intraoperative blood loss or transfusion requirement for LTME with a mean difference of 102 millilitres (MD ‐101.78; 95% CI ‐147.57 to ‐55.98; Analysis 2.6) and an odds ratio for transfusion requirement of 0.34 (95% CI 0.19 to 0.62; Analysis 2.7). Only King 2006 described a higher transfusion requirement for LTME, but a lower percentage of participants with over 100 millilitres of blood loss during surgery (27% versus 95%, P < 0.001). The overall intraoperative morbidity was described in four studies and was 11.3% for LTME versus 12.0% for OTME (OR 0.86; 95% CI 0.62 to 1.18; Analysis 2.8). There were insufficient data to compare bowel injury, haemorrhage and solid organ injury separately but individual studies did not show any differences.

Postoperative morbidity and mortality
The overall complication rate was 29.3% (LTME) and 27.5% (OTME) (OR 0.94; 95% CI 0.80 to 1.10; Analysis 3.1), with fewer wound infections and less postoperative bleeding in the LTME group (OR 0.68; 95% CI 0.50 to 0.93 (Analysis 3.2) and OR 0.30; 95% CI 0.10 to 0.93 (Analysis 3.3)). We found no differences in urinary bladder infection or retention (OR 1.23; 95% CI 0.83 to 1.81; Analysis 3.4) and pneumonia (OR 1.32; 95% CI 0.83 to 2.09; Analysis 3.5) between both groups. See Figure 4.


Forest plot of comparison: 4 Short term morbidity and mortality, outcome: 4.1 30d morbidity (total).

Forest plot of comparison: 4 Short term morbidity and mortality, outcome: 4.1 30d morbidity (total).

Ten studies described similar anastomotic leakage rate for both groups (7.7% vs 6.3% OR 1.01; 95% CI 0.73 to 1.40; Analysis 3.6), while two other trials only included abdominoperineal resection and did not have anastomotic leakage as an outcome. Consequently, the need for reoperation was 5.1% and 5.8% (OR 0.82; 95% CI 0.57 to 1.20; Analysis 3.7) in the LTME and OTME groups respectively. The anastomotic leakage rate has been corrected for participants without an anastomosis.

Data on postoperative mortality were available for 11 studies, with similar mortality rates for the two treatment groups for individual and grouped data (OR 0.81; 95% CI 0.50 to 1.32; Analysis 3.8). Four of them (Zhou 2004; Kang 2010; Liu 2010; Liang 2011) reported no 30‐day mortality in either group.

Postoperative pain
Postoperative pain can be assessed in many different ways. Common measures are a visual analogical scale (VAS) score, patient‐controlled anaesthesia (PCA) use, days of morphine use and epidural insufficiency requiring opioid use. Six studies reported results for pain score and analgesic use, and all reported lower analgesic use in the LTME group (standardised mean difference (SMD) ‐0.60; 95% CI ‐0.93 to ‐0.27; Analysis 4.1). COLOR 2 a 2013 reported on the percentage of participants using epidural, opioids or other analgesics, with less epidural use in the LTME group.Three trials reported on VAS pain scores, with a lower pain score for LTME at day one (MD ‐0.74; 95% CI ‐1.04 to ‐0.44; Analysis 4.2).

Gastrointestinal recovery and hospital stay
Length of hospital stay was given in 11 studies, and showed a reduction of two days for the LTME group (MD ‐2.16; 95% CI ‐3.22 to ‐1.10; Analysis 4.3). This is reflected in the gastrointestinal recovery rate to a faster resumption of a normal diet (MD ‐0.52; 95% CI ‐0.80 to ‐0.23; Analysis 4.4), and an earlier first bowel movement (MD ‐0.86; 95% CI ‐1.17 to ‐0.54; Analysis 4.5) in the LTME group. See Figure 5.


Forest plot of comparison: 5 Post op recovery, outcome: 5.3 Hospital stay.

Forest plot of comparison: 5 Post op recovery, outcome: 5.3 Hospital stay.

Long‐term morbidity: Incisional herniae and bowel obstruction
Only three studies reported on long‐term morbidity from incisional hernia and intestinal obstruction (MRC CLASICC a 2005; Braga 2007; Hong Kong b 2009). No statistically significant difference between OTME and LTME was seen (OR 0.84; 95% CI 0.32 to 2.21; Analysis 5.1). Intestinal obstruction occurred less frequently in the LTME group (OR 0.30; 95% CI 0.12 to 0.75; Analysis 5.2).

Quality of life: physical and sexual functioning
Four studies reported on quality of life using questionnaires (MRC CLASICC a 2005; King 2006; Braga 2007; Kang 2010). Only two reported on bladder and sexual functioning (MRC CLASICC a 2005; Kang 2010).

Of the four reporting on physical functioning, three reported significantly better functioning in the LTME group at three, six or 12 months. The MRC CLASICC a 2005 showed return to normal functioning at three months for both groups.

The reports on bladder and sexual functioning suffered from low response rates, varying from 71% overall response rate down to 10% on specific questions about sexual enjoyment and problems. Kang 2010 showed a baseline difference in sexual problems, but better sexual functioning after three months in both groups. In contrast, male sexual problems were worse three months after surgery but there was no difference between both groups. The LTME group had significantly fewer micturition, gastrointestinal and defecation problems at three months after surgery.

MRC CLASICC a 2005 and MRC CLASICC b 2005 both reported on participants in the CLASICC trial, but used different populations, questionnaires and time points. MRC CLASICC b 2005 showed worse sexual functioning after LTME, specifically for erectile dysfunction, but none were statistically significant. No differences in sexual interest, activity and enjoyment were seen at any time point, although for women there was a significant decrease compared to the preoperative baseline for both groups.

Immune response
Five studies described some short‐term differences in immune response (MRC CLASICC c 2001; Hong Kong d 2003; Zhou 2007; Hong Kong c 2000; COLOR 2 b 2011). They all reported different parameters, including C‐reactive protein (CRP), white blood cell count (WBC) and Interleukin‐6 (IL‐6). Two studies reported on B‐cell, T‐cell, cortisol and natural‐killer cell (NK‐cell) levels.

MRC CLASICC c 2001 had the largest population (n = 161), but did not show any differences in T‐cell, B‐cell and NK‐cell levels at day three. Zhou 2007 included 71 participants and the three other studies around 40 participants each. Hong Kong c 2000 showed higher levels of IL‐6 and CRP, with a peak for IL‐6 at two hours (P < 0.001) and CRP at 48 hours (P < 0.01) in the OTME group. The same results were shown by Zhou 2007, but they were measured at one and three days with a difference for IL‐6 at day one and for CRP at day one and three for the OTME group. Cortisol levels and WBC were also higher in the OTME group at day one. COLOR 2 b 2011 expressed results only as ratios compared to preoperative values and showed less increase in IL‐6 level at two hours postoperatively in the LTME group. Cortisol, WBC and CRP did not show any differences at 2, 24 and 72 hours. Finally, Hong Kong d 2003 did not show any differences at days one and three for WBC, NK‐cell, T‐cell and B‐cell levels, but for T‐cell and B‐cell levels there was less suppression in the LTME group at day eight.

Costs
An analysis of costs was included in five studies (Hong Kong a 2004; MRC CLASICC a 2005; King 2006; Braga 2007; Ng 2008). Data were too heterogeneous to be included in a meta‐analysis. Braga 2007 only reported the difference in costs in which the benefits of LTME could not compensate for the additional operating room charges, with a mean difference of USD 351 more for LTME. The four other studies calculated the costs per participant randomised. King 2006 and MRC CLASICC a 2005 reported the median direct and indirect costs for LTME. King 2006 reported the costs at GBP 6344 for LTME and GBP 6786 for OTME resulting in a saving of GBP 353 for LTME while being the only study in this analysis that used a fast‐track programme. MRC CLASICC a 2005 reported the opposite, with GBP 8259 for LTME and GBP 7820 for OTME, resulting in GBP 439 higher costs for LTME. Neither result achieved a statistically significant difference. Hong Kong a 2004 and Ng 2008 reported only the direct costs, with means of USD 9297 and USD 9588 for LTME and USD 7148 and USD 7517 for OTME with a significant difference of about USD 2000 in favour of OTME.

Discussion

Summary of main results

Nine studies (n = 1877) reported on at least one of the long‐term survival or recurrence outcomes and the meta‐analyses as well as the separate studies showed similar long‐term survival and recurrence rates for laparoscopic and open total mesorectal excision.

We found a mean difference in hospital stay of two days, with individual studies reporting a 0.5‐ to 5‐day difference in favour of LTME. Schwenk 2005 found comparable results for colon cancer with a mean difference of 1.5 days in favour of the laparoscopic group. Seven studies standardised their postoperative protocol, but only two implemented an enhanced recovery programme.

Overall completeness and applicability of evidence

Benefits of laparoscopic surgery are attributable to causing less surgical trauma to the patient, which has a positive effect on surgery‐induced immunosuppression. This can be demonstrated by taking measurements after surgery, with different peak moments for several parameters. The included studies in this review did not take measurements at the same time point, which may explain why not all of them could show differences in similar parameters. Reduced immunosuppression could be related to a lower complication rate and to shorter hospital stay, and may reduce development of postoperative metastasis (although this has yet to be shown in a randomised trial (Hogan 2011)).

The included RCTs include studies and subgroups of patients with low, mid and high rectal cancer and both APR and anterior resections with and without anastomosis. These differences can affect outcomes, especially the various techniques for low rectal resections can influence the circumferential margins and therefore local recurrences and survival. Lack of reporting of the CRM is an important issue with only eight out of fourteen studies describing this outcome. The number of retrieved lymph nodes is described by eleven studies but is more dependent on difference in high and low vascular ligations of the mesentery than on open or laparoscopic surgery. (Kessler 2013)

With a mean age between 44 and 72, three studies including T4 carcinoma and six offering neoadjuvant treatment in selected cases and a tumour localisation between 15 cm and the anal verge, there is a fair amount of heterogeneity among the included studies. Especially the early and smaller studies included a younger and healthier study population compared to the average rectal cancer patient. Of the four ongoing trials (n = 470 to 1100 participants), three (Kang 2010; ACTRN12609000663257; NCT00726622) require neoadjuvant treatment for selected stages of rectal cancer and the fourth (COLOR 2 a 2013) stratifies the randomisation for neoadjuvant treatment. This might influence both long term and short term outcomes as only six offered neoadjuvant therapy in this review. All ongoing studies include abdominoperineal resections as well as (low) anterior resections, but the maximum distance from the anal verge varies between 9 cm (Kang 2010), 12 cm (NCT00726622) and 15 cm (ACTRN12609000663257; COLOR 2 a 2013).

Another important difference between the included RCTs and current practise are the fast track recovery programmes such as the enhanced recovery after surgery (ERAS) programme. Only two included studies describe an enhanced recovery programme (King 2006; Lujan 2009). COLOR 2 a 2013 referred to local protocols, whereas the other two ongoing studies do not describe their postoperative protocol in the online summary. The LAFA trial (Vlug 2011) showed laparoscopic surgery in combination with fast track recovery resulted in the fastest recovery and hospital discharge compared to regular care and open surgery.

Quality of the evidence

Since 1998, 14 RCTs have been published to answer the question whether LTME results in better short‐term results and at least equal long‐term oncological results. The quality of these studies varied extensively, as did the number of included participants. Although the total mesorectal excision principle has been established since 1986, treatment protocols have changed. Surgeons gained more experience in laparoscopic colon and rectal cancer surgery, fast‐track protocols were introduced and neoadjuvant treatment became a standard of care in a proportion of cases. All of these factors are able to influence the long‐term results of these trials; however they should influence both the laparoscopic and open groups equally, except for the learning curve for laparoscopic procedures.

The quality of the evidence for the most important outcomes was moderate (summary of findings Table for the main comparison). This means that further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. The main reason is the imprecision of the confidence intervals, they allow for a variability in odds ratios up to 40% on both sides, which contributes to an absolute increase or decrease of 6% in disease free survival, 5% in overall survival, 2% in local recurrences, and 3% in 30‐day morbidity. The COLOR 2 a 2013 trial used a 5% margin for local recurrences for their non‐inferiority design, therefore our results remain within those limits.

Potential biases in the review process

Publication bias is a threat for any systematic review. We believe we have missed no important randomised controlled trials after screening reference lists of included trials and other relevant studies and reviews, in addition to the extensive systematic searching of electronic databases and trials registers. We have described all registered ongoing trials.

In the current literature regarding learning curves in laparoscopic colorectal surgery, a wide range of numbers of procedures is reported until a flat curve is achieve, ranging from 11 to 15 colectomies (Simons 1995), 30 colorectal resections (Schlachta 2001) and 60 to 65 colectomies (Tekkis 2005). For the open total mesorectal excision (OTME) technique, the cut‐off point for percentage of clear resection margins is defined as around 50 procedures (Oh 2011). This suggests that only the surgeons in Kang 2010 are assumed to have had sufficient experience for a good laparoscopic resection and results may further improve over time.

Agreements and disagreements with other studies or reviews

A meta‐analysis of RCTs on laparoscopic and open colorectal surgery (Sammour 2011) has shown a higher intraoperative complication rate for laparoscopic surgery of 6.3% versus 3.9% for open surgery (OR)1.55, 95%CI 1.12 to 2.15). The rate of bowel perforations was 2.1% versus 0.9% (OR 2.28, 95% CI 1.27 to 4.10) across 3018 participants. These differences had limited effect on the outcome, with an average postoperative complication rate of 28%. Compared to the intraoperative complication rate of 11.3% in LTME versus 12.0% in OTME in four included studies in our review (n = 1618), we cannot confirm these previously reported complication rates for LTME.

The results of this review confirm what other colorectal and rectal trials have suggested: short‐term results are similar with faster recovery in the LTME group and no statistically significant differences were found in the long‐term oncological results. For rectal cancer, non‐randomised trials have suggested oncological safe resections as presented in the previous version of this review (Fleshman 1999; Feliciotti 2003; Breukink 2006). Since then, several other reviews have been published describing the same results. Aziz 2006, Gao 2006, Anderson 2008 and Poon 2009 included mainly non‐randomised trials, and Row 2010 was a literature review. Ohtani 2011, Huang 2011 and Trastulli 2012 were the first to include only randomised trials. However, Ohtani 2011 also included three non‐randomised trials, Huang 2011 included only six trials and Trastulli 2012 nine trials, whereas this systematic review was able to identify 14. In addition, the Cochrane review of laparoscopic colorectal cancer (Kuhry 2008) presented separate meta‐analyses for four included rectal cancer RCTs.

Study selection flow diagram.
Figures and Tables -
Figure 1

Study selection flow diagram.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Figures and Tables -
Figure 2

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Forest plot of comparison: 2 Survival and recurrences, outcome: 2.1 Disease free survival.
Figures and Tables -
Figure 3

Forest plot of comparison: 2 Survival and recurrences, outcome: 2.1 Disease free survival.

Forest plot of comparison: 4 Short term morbidity and mortality, outcome: 4.1 30d morbidity (total).
Figures and Tables -
Figure 4

Forest plot of comparison: 4 Short term morbidity and mortality, outcome: 4.1 30d morbidity (total).

Forest plot of comparison: 5 Post op recovery, outcome: 5.3 Hospital stay.
Figures and Tables -
Figure 5

Forest plot of comparison: 5 Post op recovery, outcome: 5.3 Hospital stay.

Comparison 1 Survival and recurrences, Outcome 1 Disease‐free survival.
Figures and Tables -
Analysis 1.1

Comparison 1 Survival and recurrences, Outcome 1 Disease‐free survival.

Comparison 1 Survival and recurrences, Outcome 2 Overall survival.
Figures and Tables -
Analysis 1.2

Comparison 1 Survival and recurrences, Outcome 2 Overall survival.

Comparison 1 Survival and recurrences, Outcome 3 Local recurrences.
Figures and Tables -
Analysis 1.3

Comparison 1 Survival and recurrences, Outcome 3 Local recurrences.

Comparison 1 Survival and recurrences, Outcome 4 Distant recurrences.
Figures and Tables -
Analysis 1.4

Comparison 1 Survival and recurrences, Outcome 4 Distant recurrences.

Comparison 1 Survival and recurrences, Outcome 5 Wound/port site metastases.
Figures and Tables -
Analysis 1.5

Comparison 1 Survival and recurrences, Outcome 5 Wound/port site metastases.

Comparison 2 Surgical data, Outcome 1 Lymph nodes retrieved.
Figures and Tables -
Analysis 2.1

Comparison 2 Surgical data, Outcome 1 Lymph nodes retrieved.

Comparison 2 Surgical data, Outcome 2 CRM positivity.
Figures and Tables -
Analysis 2.2

Comparison 2 Surgical data, Outcome 2 CRM positivity.

Comparison 2 Surgical data, Outcome 3 Duration of surgery.
Figures and Tables -
Analysis 2.3

Comparison 2 Surgical data, Outcome 3 Duration of surgery.

Comparison 2 Surgical data, Outcome 4 Incision length.
Figures and Tables -
Analysis 2.4

Comparison 2 Surgical data, Outcome 4 Incision length.

Study

Araujo 2003

0 (0/13)

Braga 2007

7.2 (6/83)

COLOR 2 a 2013

17 (121/695)

Hong Kong a 2004

23.2 (47/203)

Kang 2010

1.2 (2/170)

King 2006

7.3 (3/41)

Liang 2011

0.5 (1/169)

Liu 2010

0 (0/98)

Lujan 2009

7.9 (8/101)

MRC CLASICC a 2005

33.9 (82/242)

Ng 2008

9.8 (5/51)

Pechlivanides 2007

2.9 (1/34)

Zhou 2004

Unknown

Zhou 2007

Unknown

Figures and Tables -
Analysis 2.5

Comparison 2 Surgical data, Outcome 5 Conversion rate.

Comparison 2 Surgical data, Outcome 6 Blood loss.
Figures and Tables -
Analysis 2.6

Comparison 2 Surgical data, Outcome 6 Blood loss.

Comparison 2 Surgical data, Outcome 7 Transfusion requirement.
Figures and Tables -
Analysis 2.7

Comparison 2 Surgical data, Outcome 7 Transfusion requirement.

Comparison 2 Surgical data, Outcome 8 Intraoperative morbidity.
Figures and Tables -
Analysis 2.8

Comparison 2 Surgical data, Outcome 8 Intraoperative morbidity.

Comparison 3 Short‐term morbidity and mortality, Outcome 1 30‐day morbidity (total).
Figures and Tables -
Analysis 3.1

Comparison 3 Short‐term morbidity and mortality, Outcome 1 30‐day morbidity (total).

Comparison 3 Short‐term morbidity and mortality, Outcome 2 Wound infection.
Figures and Tables -
Analysis 3.2

Comparison 3 Short‐term morbidity and mortality, Outcome 2 Wound infection.

Comparison 3 Short‐term morbidity and mortality, Outcome 3 Bleeding.
Figures and Tables -
Analysis 3.3

Comparison 3 Short‐term morbidity and mortality, Outcome 3 Bleeding.

Comparison 3 Short‐term morbidity and mortality, Outcome 4 Urinary complications.
Figures and Tables -
Analysis 3.4

Comparison 3 Short‐term morbidity and mortality, Outcome 4 Urinary complications.

Comparison 3 Short‐term morbidity and mortality, Outcome 5 Pneumonia.
Figures and Tables -
Analysis 3.5

Comparison 3 Short‐term morbidity and mortality, Outcome 5 Pneumonia.

Comparison 3 Short‐term morbidity and mortality, Outcome 6 Anastomotic leakage.
Figures and Tables -
Analysis 3.6

Comparison 3 Short‐term morbidity and mortality, Outcome 6 Anastomotic leakage.

Comparison 3 Short‐term morbidity and mortality, Outcome 7 Need for reoperation.
Figures and Tables -
Analysis 3.7

Comparison 3 Short‐term morbidity and mortality, Outcome 7 Need for reoperation.

Comparison 3 Short‐term morbidity and mortality, Outcome 8 30‐day mortality.
Figures and Tables -
Analysis 3.8

Comparison 3 Short‐term morbidity and mortality, Outcome 8 30‐day mortality.

Comparison 4 Postoperative recovery, Outcome 1 Analgesia use (number of doses).
Figures and Tables -
Analysis 4.1

Comparison 4 Postoperative recovery, Outcome 1 Analgesia use (number of doses).

Comparison 4 Postoperative recovery, Outcome 2 Day 1 pain score (VAS).
Figures and Tables -
Analysis 4.2

Comparison 4 Postoperative recovery, Outcome 2 Day 1 pain score (VAS).

Comparison 4 Postoperative recovery, Outcome 3 Hospital stay (days).
Figures and Tables -
Analysis 4.3

Comparison 4 Postoperative recovery, Outcome 3 Hospital stay (days).

Comparison 4 Postoperative recovery, Outcome 4 Time to normal diet (days).
Figures and Tables -
Analysis 4.4

Comparison 4 Postoperative recovery, Outcome 4 Time to normal diet (days).

Comparison 4 Postoperative recovery, Outcome 5 Time to first defecation (days).
Figures and Tables -
Analysis 4.5

Comparison 4 Postoperative recovery, Outcome 5 Time to first defecation (days).

Comparison 5 Long term morbidity, Outcome 1 Incisional hernia.
Figures and Tables -
Analysis 5.1

Comparison 5 Long term morbidity, Outcome 1 Incisional hernia.

Comparison 5 Long term morbidity, Outcome 2 Intestinal obstruction.
Figures and Tables -
Analysis 5.2

Comparison 5 Long term morbidity, Outcome 2 Intestinal obstruction.

Summary of findings for the main comparison. Laparoscopic versus open total mesorectal excision for rectal cancer

Laparoscopic versus open total mesorectal excision (TME) for rectal cancer

Patient or population: people with Rectal Cancer
Settings:
Intervention: Laparoscopic TME
Comparison: Open TME

Outcomes

Illustrative comparative risks* (95% CI)

Relative effect
(95% CI)

No of Participants
(studies)

Quality of the evidence
(GRADE)

Assumed risk

Corresponding risk

Open TME

Laparoscopic TME

Disease‐free survival at 5 years

718 per 1000

722 per 1000
(659 to 778)

OR 1.02
(0.76 to 1.38)

943
(4 studies)

⊕⊕⊕⊝
moderate1

Overall survival at 5 years

679 per 1000

709 per 1000
(648 to 763)

OR 1.15
(0.87 to 1.52)

987
(4 studies)

⊕⊕⊕⊝
moderate2

Local recurrences

54 per 1000

48 per 1000
(31 to 73)

OR 0.89
(0.57 to 1.39)

1538
(8 studies)

⊕⊕⊕⊝
moderate3

Lymph nodes retrieved

The mean number of lymph nodes retrieved in the intervention groups was
0.43 lower
(1.13 lower to 0.26 higher)

3682
(11 studies)

⊕⊕⊕⊕
high

CRM positivity

61 per 1000

60 per 1000
(44 to 83)

OR 0.99
(0.71 to 1.4)

2313
(8 studies)

⊕⊕⊕⊝
moderate4

30‐day morbidity (total)

275 per 1000

263 per 1000
(233 to 295)

OR 0.94
(0.8 to 1.1)

3397
(11 studies)

⊕⊕⊕⊝
moderate5

Hospital stay (days)

The mean length of hospital stay in the intervention groups was
2.16 days shorter
(3.22 to 1.1 days shorter)

3084
(11 studies)

⊕⊕⊕⊝
moderate6

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; OR: Odds ratio;

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

1Statistical inaccuracy with wide confidence interval at both sides

2Statistical inaccuracy with wide confidence interval at both sides, but a tendency for a higher overall survival for LTME

3Statistical inaccuracy with wide confidence interval at both sides, but a tendency for a lower recurrence rate for LTME

4Only 8 studies provided data on CRM positivity
5Definition of overall morbidity varied or was unclear
6Length of hospital stay depends on postoperative protocols and implementation of enhanced recovery programmes

Figures and Tables -
Summary of findings for the main comparison. Laparoscopic versus open total mesorectal excision for rectal cancer
Table 1. Reported outcomes

Study ID

n

Long‐term survival

30‐day mortality

30‐day morbidity

Long‐term morbidity

Lymphnodes

Gastrointestinal recovery

Pain

Bleeding

Length of hospital stay

Immune response

Quality of life

Cost

Araujo 2003

28

+

+

+

+

Braga 2007

168

5y/3y

+

+

+

+

+

+

+

+

+

COLOR 2 a 2013

1044

+

+

+

+

+

+

+

COLOR 2 b 2011

40

+

+

+

+

+

+

Hong Kong a 2004

403

5y

+

+

+

+

+

+

+

+

Hong Kong b 2009

153

10y

+

+

Hong Kong c 2000

34

+

Hong Kong d 2003

40

+

Kang 2010

340

+

+

+

+

+

+

+

+

King 2006

19

+

+

+

+

+

+

Liang 2011

343

3y

+

+

+

+

+

Liu 2010

186

+

+

+

+

+

Lujan 2009

204

5y

+

+

+

+

+

+

MRC CLASICC a 2005

381

10y/5y/3y

+

+

+

+

+

+

MRC CLASICC b 2005

148

+

MRC CLASICC c 2001

236

+

Ng 2008

99

5y

+

+

+

+

+

+

+

+

Pechlivanides 2007

73

+

Zhou 2004

171

+

+

+

+

+

Zhou 2007

71

+

Figures and Tables -
Table 1. Reported outcomes
Comparison 1. Survival and recurrences

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

1 Disease‐free survival Show forest plot

5

Odds Ratio (M‐H, Fixed, 95% CI)

Subtotals only

1.1 10‐year

1

130

Odds Ratio (M‐H, Fixed, 95% CI)

1.25 [0.51, 3.06]

1.2 5‐year

4

943

Odds Ratio (M‐H, Fixed, 95% CI)

1.02 [0.76, 1.38]

1.3 3‐year

1

326

Odds Ratio (M‐H, Fixed, 95% CI)

1.08 [0.67, 1.74]

2 Overall survival Show forest plot

6

Odds Ratio (M‐H, Fixed, 95% CI)

Subtotals only

2.1 10‐year

2

534

Odds Ratio (M‐H, Fixed, 95% CI)

1.15 [0.80, 1.65]

2.2 5‐year

4

987

Odds Ratio (M‐H, Fixed, 95% CI)

1.15 [0.87, 1.52]

2.3 3‐year

2

682

Odds Ratio (M‐H, Fixed, 95% CI)

1.00 [0.70, 1.42]

3 Local recurrences Show forest plot

8

1538

Odds Ratio (M‐H, Fixed, 95% CI)

0.89 [0.57, 1.39]

3.1 5‐year

5

963

Odds Ratio (M‐H, Fixed, 95% CI)

0.94 [0.49, 1.81]

3.2 3‐year

3

575

Odds Ratio (M‐H, Fixed, 95% CI)

0.84 [0.46, 1.56]

4 Distant recurrences Show forest plot

6

1341

Odds Ratio (M‐H, Fixed, 95% CI)

0.96 [0.70, 1.32]

5 Wound/port site metastases Show forest plot

7

2130

Odds Ratio (M‐H, Fixed, 95% CI)

2.76 [0.75, 10.20]

Figures and Tables -
Comparison 1. Survival and recurrences
Comparison 2. Surgical data

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

1 Lymph nodes retrieved Show forest plot

11

3682

Mean Difference (IV, Random, 95% CI)

‐0.43 [‐1.13, 0.26]

2 CRM positivity Show forest plot

8

2313

Odds Ratio (M‐H, Fixed, 95% CI)

0.99 [0.71, 1.40]

3 Duration of surgery Show forest plot

12

3840

Mean Difference (IV, Random, 95% CI)

37.48 [27.80, 47.15]

4 Incision length Show forest plot

4

1488

Mean Difference (IV, Random, 95% CI)

‐12.83 [‐14.87, ‐10.80]

5 Conversion rate Show forest plot

Other data

No numeric data

6 Blood loss Show forest plot

8

2615

Mean Difference (IV, Random, 95% CI)

‐101.78 [‐147.57, ‐55.98]

7 Transfusion requirement Show forest plot

5

939

Odds Ratio (M‐H, Fixed, 95% CI)

0.34 [0.19, 0.62]

8 Intraoperative morbidity Show forest plot

4

1618

Odds Ratio (M‐H, Fixed, 95% CI)

0.86 [0.62, 1.18]

Figures and Tables -
Comparison 2. Surgical data
Comparison 3. Short‐term morbidity and mortality

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

1 30‐day morbidity (total) Show forest plot

11

3397

Odds Ratio (M‐H, Fixed, 95% CI)

0.94 [0.80, 1.10]

2 Wound infection Show forest plot

10

3337

Odds Ratio (M‐H, Fixed, 95% CI)

0.68 [0.50, 0.93]

3 Bleeding Show forest plot

5

1181

Odds Ratio (M‐H, Fixed, 95% CI)

0.30 [0.10, 0.93]

4 Urinary complications Show forest plot

8

1756

Odds Ratio (M‐H, Fixed, 95% CI)

1.23 [0.83, 1.81]

5 Pneumonia Show forest plot

8

2668

Odds Ratio (M‐H, Fixed, 95% CI)

1.32 [0.83, 2.09]

6 Anastomotic leakage Show forest plot

10

2505

Odds Ratio (M‐H, Fixed, 95% CI)

1.01 [0.73, 1.40]

7 Need for reoperation Show forest plot

7

2316

Odds Ratio (M‐H, Fixed, 95% CI)

0.82 [0.57, 1.20]

8 30‐day mortality Show forest plot

11

3812

Odds Ratio (M‐H, Fixed, 95% CI)

0.81 [0.50, 1.32]

Figures and Tables -
Comparison 3. Short‐term morbidity and mortality
Comparison 4. Postoperative recovery

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

1 Analgesia use (number of doses) Show forest plot

5

1199

Std. Mean Difference (IV, Random, 95% CI)

‐0.60 [‐0.93, ‐0.27]

2 Day 1 pain score (VAS) Show forest plot

3

776

Mean Difference (IV, Fixed, 95% CI)

‐0.74 [‐1.04, ‐0.44]

3 Hospital stay (days) Show forest plot

11

3084

Mean Difference (IV, Random, 95% CI)

‐2.16 [‐3.22, ‐1.10]

4 Time to normal diet (days) Show forest plot

8

2109

Mean Difference (IV, Random, 95% CI)

‐0.52 [‐0.80, ‐0.23]

5 Time to first defecation (days) Show forest plot

8

2893

Mean Difference (IV, Random, 95% CI)

‐0.86 [‐1.17, ‐0.54]

Figures and Tables -
Comparison 4. Postoperative recovery
Comparison 5. Long term morbidity

Outcome or subgroup title

No. of studies

No. of participants

Statistical method

Effect size

1 Incisional hernia Show forest plot

3

508

Odds Ratio (M‐H, Random, 95% CI)

0.84 [0.32, 2.21]

2 Intestinal obstruction Show forest plot

3

508

Odds Ratio (M‐H, Fixed, 95% CI)

0.30 [0.12, 0.75]

Figures and Tables -
Comparison 5. Long term morbidity